Why We Failed to Win a Decisive Victory in Afghanistan
Shifting political allegiances, not smashing enemies, should have been the goal. And that holds true for the campaign against the Islamic State as well.
There’s been a great debate over on Tom Ricks’s Best Defense blog in response to Jim Gourley’s question in relation to the Afghanistan campaign: “Why did we fail to render our enemies — those people who actively participated in open hostility against our forces — powerless?” Perhaps not surprisingly, many of the responses have taken the view that we have, indeed, lost in Afghanistan.
I take issue, however, with the starting assumption that “rendering our enemies powerless” should be the standard by which we evaluate the success of military action in Afghanistan, or lack thereof. I think the assumption clouds the analysis of both of Afghanistan and the conflict against the so-called Islamic State.
So here’s my answer.
War has two meanings. The first is a descriptive sense, in that war describes a situation above a certain threshold of violence, and therefore includes conflicts like Iraq and Afghanistan. The second is an instrumental sense, meaning a particular way in which force is used to achieve a political goal. The default understanding in Western militaries of war in the instrumental sense is still Clausewitzian. Consider the opening page of On War: “we must render the enemy powerless; and that, in theory, is the true aim of warfare.”
But when Clausewitz wrote that, he meant a very specific kind of war, namely the use of force in the Napoleonic context that was about decisive battle and unconditional surrender. When he wrote about more limited uses of force elsewhere in On War, he said that the more political considerations displaced military considerations, the more his theory of Napoleonic absolute war had to be adapted to account for the fact that outcomes short of the enemy’s overthrow might be the most realistic policy aim.
To my mind, “rendering our enemies powerless” is too narrow a concept of success to analyze the Afghan campaign, because “powerless” assumes that only decisive defeat of an enemy counts as success; anything else means failure. That excludes the successful use of military force in circumstances in which a decisive outcome is not realistic.
In Afghanistan, as in Iraq, when the conventional phase was over and the mission became indistinguishable from enforcing the writ of a relatively corrupt government over disillusioned parts of its own population, the notion that a decisive outcome was even available is illusory: first, because that task is endless — as it’s about changing people’s political affiliations, which are liable to evolve (as we have seen quite spectacularly in Iraq since the surge); second, because there was not a single coherent enemy force to be rendered powerless in the first place.
In Afghanistan, there are tranches of the enemy who can and should be decisively defeated, like hard-core jihadi cells. Yet many “insurgents” are actually criminals — to whom the concept of decisive defeat is inappropriate, as criminality will always be there. Moreover, many of these criminals pursue local goals, and have no interest in marching on the Kabul government, which they’d be happily living off of parasitically. Other “insurgents” are locals with legitimate grievances against a corrupt government, against whom the concept of rendering them powerless is not just inappropriate but positively damaging: against them, success should be thought of in terms of empowering them.
By analogy, take the 2007-2008 Iraq surge. Sure, there were hard-core al Qaeda cells, the fight against whom by coalition Special Forces can legitimately be analyzed in terms of decisive battlefield outcomes. However, most of the insurgents were Sunnis alienated from Baghdad whom the surge empowered by resetting their relationship with the central government, thus giving them some of what they wanted. (And it is indeed their subsequent disempowerment, by the Shiite former Prime Minister Nouri al-Maliki — who was precisely trying to achieve his political ends by rendering them powerless — that has led to the current chaos.)
In short, the idea that success in insurgencies necessarily means rendering all enemies powerless often mischaracterizes as failures outcomes that have actually proven successful in bringing peace.
The broader issue here, which is still live in light of how we approach other contemporary conflicts, is whether or not counterinsurgency was the wrong operational approach because it didn’t deliver decisive results.
Take the counterfactual argument that we could have achieved a decisive result if we had left Afghanistan in 2002, having conventionally defeated the Taliban and al Qaeda and routing them from Kabul, treating any further security issues as the Afghan government’s responsibility. OK, fine. But what if we had had to go back in later — would that have changed the evaluation of the decisiveness of the outcome achieved in 2002? Maybe, maybe not; the answer depends entirely on the scope of the conflict to which you attach the concept of success. Is it sufficient just to think about military success as rendering an enemy powerless, or is military success really about the political conditions the use of force enables?
It seems to me that you have to include the achievement of policy goals — not just the evaluation of how an enemy comes off from the fight — in the evaluation of decisiveness of military outcome, or else the use of force becomes disconnected from its political purpose.
On that basis, to those of you who say we should have had a counterterrorism (CT) rather than a counterinsurgency (COIN) approach in the post-2002 phase of Afghanistan, I say whatever the merits of your argument in terms of its effect on the enemy, a CT approach would have been incompatible with the policy goals: your real issue is with foreign policy, not COIN.
Let’s briefly unpack that. A CT approach, given that insurgents in Afghanistan generally refuse to fight us conventionally, realistically would have meant working through militias, who are plainly not democratic. If your policy goals therefore require democratic reform you can’t have a CT approach that empowers militias (democracy, women’s rights, and anti-narcotics, for example, were still touted as coalition ambitions well after the policy goal was ostensibly reduced to “defeat, disrupt and dismantle” al Qaeda in 2009).
Was the COIN campaign necessitated by idealistic policy goals ever realistically going to be able to decisively deliver? No. A decisive outcome to the Afghan campaign was never on the table in the first place given the policy goals. The campaign was never reducible to a zero-sum outcome of victory and defeat, given that the long-term security of Afghanistan depended on issues beyond defeating the enemy, and the enemy himself was not a coherent entity against whom a single concept of defeat, in the sense of rendering it powerless, could even be applied.
Of course, you can ignore all that and force a military outcome on an insurgent enemy, by killing them alongside the civilian population amongst whom they hide, which is what Sri Lanka did in its 2009 civil war. It worked, but it was brutal, immoral, and did not address the underlying political grievances necessary for a long-term peace. Western armies rightly don’t do that.
If you can’t distill the Afghan campaign to a binary victory or defeat model, how then do you analyze it? To my mind, any non-Sri Lankan style counterinsurgency approach (and I don’t just mean heavy footprint COIN like in Afghanistan, but all COIN, including the use of small teams of advisors with local units in other contemporary conflicts) can be understood as an effort to persuade a range of constituencies, including the various parts of the enemy, to subscribe to a given narrative, or political story, if you like. At the very core in Afghanistan, that narrative was “don’t fight the Afghan government.”
Some parts of the enemy are never going to come across to our narrative, so you do need to render them powerless, which means using violent force against them. However, to take on all insurgents through violent force is inappropriate — both from a moral and a resource point of view. The failure to distinguish between different types of insurgents was a key problem early on in the Afghan campaign: we treated everyone who shot at us as hard-core “Taliban,” as if they were a single enemy, and ended up swimming upstream against a large insurgency.
Many insurgents, like narco-factions in Helmand, were pushed away from the Afghan government’s writ because of the anti-narcotics policy goal, without which southern Afghanistan would have been far less violent. It’s not surprising that, in an economy based on opium, this inflated the size of the insurgency with fighters who would not otherwise have taken up arms. Take, for example, Sher Mohammad Akhundzada, who flipped his faction to fight the Afghan government having been sacked as the provincial governor of Helmand for narcotics offenses.
Yes, I know opium funds the Taliban, but it also funds most of the economy of southern Afghanistan; so please don’t come back to me with the argument that somehow destroying opium specifically hurt the Taliban rather than alienating the whole region. The net effect was to massively help the Taliban. The policy also shifted onto us the burden of responsibility for finding an alternative economy for southern Afghanistan, another avoidable gargantuan task. At least democracy, rule of law, and women’s rights are honorable ambitions, even if they were unrealistic in many parts of Afghanistan. The anti-drugs aim, however, was an utterly ruinous, delusionary, and unnecessary foreign-policy goal that to my mind is the factor that, above all, unhinged our campaign in the south.
We were more effective later in the campaign when our understanding was far more nuanced and we stopped being such ideologues, but the earlier damage had already been done: Alienation of a constituency, whether deliberate, reckless, or accidental, is much easier than regaining their affiliation.
The bottom line is that, like in Iraq, there was scope to use aggressive battlefield action in Afghanistan to defeat tranches of the enemy like hard-core jihadi cells, not your Pashtun peasant with legitimate grievances against an abusive government. But the broader campaign, given the policy goals, was not reducible to a decisive battlefield outcome, as it was conceptually indistinguishable from extending the writ of the Afghan government, which simply meant (and still means) getting people to change their political affiliations. And since politics doesn’t end, we have entered, unsurprisingly, a long war with no clear end point, there being no clear demarcation between a peaceful “political” and wartime “military” phase of the Afghan conflict.
One soon ends up talking about “stability,” not victory — which is exactly how the language in which outcomes were described evolved through the Afghan campaign. What do you call the use of force in such a context if you are not calling it war in the instrumental sense, even if it’s war in a descriptive sense? Personally, I call it armed politics.
So let me go back to the basic conceptual reason I laid out for why we mis-analyze Afghanistan if we overly privilege the idea that it’s a war. Just because it’s descriptively a war, that doesn’t mean victory and defeat are always going to be on the table in a decisive sense. However, it was precisely the attempt to shoehorn the Afghan conflict into an instrumental model of war in which the aim is by default to render an enemy powerless that led us to treat all insurgents as part of one enemy who could be decisively defeated on the battlefield. That conceptual mistake ultimately expanded the insurgency, until we reversed out of it later in the campaign, and realized the Taliban were not a monolith and shouldn’t be fought as such. It is also this conceptual confusion that creates a false debate about why we “lost,” when Afghanistan is a conflict in which an unsatisfying outcome is probably as much as we could hope for — given how unrealistic the policy goals were. I don’t think there’s much more the military could have done given the foreign-policy context. The COIN vs. CT debate is a red herring: it’s really a proxy for differences about foreign policy.
The practical takeaway for practitioners: Taking on an insurgency is as much about the control of political as geographical space. So speaking to a political communications consultant who knows the local politics would probably be a better use of time that spending sleepless nights in the maneuver warfare library trying to work out how to converge on the decisive point and smash your enemy.
The practical takeaway for policymakers: Most contemporary conflicts, being networked conflicts of our information age, are not simply two-way fights which Western forces can win through decisive victories, not least because no sensible enemy will take on the U.S. military in a conventional fight. So if the United States genuinely only wants to fight wars it can “win” in the decisive sense, that means in reality staying out of the vast majority of contemporary conflicts and taking risk on the security implications of so doing. Realistically, that’s not going to happen, which then engages the issue of how you characterize the use of force in a conflict in which decisive battlefield outcome isn’t on the table. I suggest that armed politics is a more appropriate concept.
I don’t think it’s helpful to present the fight against al Qaeda or the Islamic State as “wars” because it condemns us to losing on our own terms. We are never going to defeat al Qaeda, or IS for that matter, through an exclusively battlefield mechanism, given they are essentially ideologies. The Cold War is a good analogy — if it was actually a war, then we would not have used an adjective that made clear that war alone was not the right way to think about that conflict, and appropriately implied that a decisive battlefield outcome was not available (at least outside the nuclear option, which is in a sense analogous to Sri Lankan-style COIN). We can achieve successes against IS, and deal with parts of the problem through aggressive battlefield action, but that plainly doesn’t mean a decisive battlefield outcome is available in the conflict as a whole. Indeed, the struggle is better understood as a long-term political effort — with elements of armed politics and domestic law enforcement — to get a wide range of constituencies not to support IS. Successes in this conflict probably won’t be particularly satisfying (think of the Shiite militias we seem to be empowering in Iraq, for example).
Finally, let’s go back to the original question: did we succeed or fail in Afghanistan? We don’t know yet, because it’s not over. Rather than speculate about how things might turn out, take the current situation: a continuing insurgency across the Pashtun provinces and parts of the north. It’s bad, but it does not present an existential threat to the state, so long as the Afghan Security Forces don’t crumble, which in turn seems to depend on outside support.
If you take security, and particularly Western security, as the basic goal, that state of affairs could be viewed as a moderate success for now, but the same result could likely have been achieved at much less cost, earlier on, with less ambitious policy goals. The anti-narcotics effort, needless to say, has been a total failure. On democracy and social goals, I don’t feel I have the authority to evaluate the quality of an Afghan girls’ life, for example, under a deeply corrupt Afghan state. But it’s clear the goals of a Westernized Afghanistan were completely unrealistic.
Do I think about the Afghan conflict with a deep sense of grief, given how policy goals untethered to reality expanded a fight in which many friends died? Yes. But I unless the Taliban march into Kabul, I don’t think we’ve lost.
I would like to thank Jim Gourley, conversation with whom was very helpful in the development of this piece.
Emile Simpson is a former British Army officer and the author of War From the Ground Up: Twenty-First Century Combat as Politics.